Image Captioning on DOCCI 500 (test)
64.31CAPTURESFT + Ours
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SFT + Oursbase_model=Qwen2-VL-7B SFT, method=CIM2026.03 | 64.31 | 45.16 | 24.88 | — | — | 73.87 | — | — | 58.68 | 36.32 | |
| 7B-OursBase Model=Qwen2.5-VL, Method=CIM2026.03 | 63.46 | 39.8 | 22.72 | 79.28 | 66.97 | 71.88 | 73.52 | 50.45 | 59.08 | 34.7 | |
| CIMBase Model=Qwen2.5-VL-7B2026.03 | 63.46 | — | — | 79.28 | 66.97 | 71.88 | 73.52 | 50.45 | 59.08 | 34.7 | |
| 7B-OursBase Model=Qwen2-VL, Method=CIM2026.03 | 63.12 | 37.21 | 21.53 | 81.99 | 64.61 | 71.43 | 74.49 | 50.19 | 59.18 | 32.12 | |
| CIMBase Model=Qwen2-VL-7B2026.03 | 63.12 | — | — | 81.99 | 64.61 | 71.43 | 74.49 | 50.19 | 59.18 | 32.12 | |
| 8B-OursBase Model=InternVL3, Method=CIM2026.03 | 62.18 | 28.25 | 18.54 | 82.73 | 62.86 | 70.47 | 76.22 | 49.58 | 59.26 | 30.39 | |
| CIMBase Model=InternVL3-8B2026.03 | 62.18 | — | — | 82.73 | 62.86 | 70.47 | 76.22 | 49.58 | 59.26 | 30.39 | |
| 72B-BaseBase Model=Qwen2.5-VL2026.03 | 61.09 | 33.72 | 19.01 | 82.17 | 62.92 | 70.35 | 72.37 | 46.18 | 55.46 | 30.71 | |
| 76B-BaseBase Model=InternVL22026.03 | 61.09 | 38.69 | 21.21 | 81.42 | 64.77 | 71.19 | 70.9 | 46.8 | 55.57 | 29.67 | |
| Claude-3.72026.03 | 61.02 | 41.36 | 20.17 | — | — | 69.48 | — | — | 55.57 | 28.78 | |
| 8B-OursBase Model=InternVL2, Method=CIM2026.03 | 60.82 | 32.01 | 19.49 | 82.16 | 60.35 | 68.66 | 74.16 | 47.51 | 56.98 | 26.36 | |
| CIMBase Model=InternVL2-8B2026.03 | 60.82 | — | — | 82.16 | 60.35 | 68.66 | 74.16 | 47.51 | 56.98 | 26.36 | |
| 8B-OursBase Model=InternVL2.5, Method=CIM2026.03 | 60.75 | 30.89 | 17.84 | 82.17 | 62.28 | 69.99 | 72.96 | 46.64 | 56.07 | 28.9 | |
| CIMBase Model=InternVL2.5-8B2026.03 | 60.75 | — | — | 82.17 | 62.28 | 69.99 | 72.96 | 46.64 | 56.07 | 28.9 | |
| Gemini-1.52026.03 | 60.34 | 24.7 | 16.25 | — | — | 68.17 | — | — | 55.69 | 28.9 | |
| 72B-BaseBase Model=Qwen2-VL2026.03 | 59.56 | 30.79 | 17.57 | 84.02 | 60.21 | 69.14 | 71.63 | 44.31 | 53.92 | 27.21 | |
| 78B-BaseBase Model=InternVL2.52026.03 | 59.37 | 23.6 | 15.79 | 85.78 | 59.12 | 68.8 | 75.12 | 45.5 | 55.74 | 26.92 | |
| 8B-BaseBase Model=InternVL22026.03 | 58.83 | 31.49 | 17.97 | 81.49 | 59.54 | 67.72 | 70.84 | 44.24 | 53.51 | 22.65 | |
| BaseBase Model=InternVL2-8B2026.03 | 58.83 | — | — | 81.49 | 59.54 | 67.72 | 70.84 | 44.24 | 53.51 | 22.65 | |
| 8B-BaseBase Model=InternVL2.52026.03 | 58.64 | 24.23 | 15.29 | 84.11 | 58.07 | 67.62 | 73.48 | 44.79 | 54.69 | 24.63 | |
| BaseBase Model=InternVL2.5-8B2026.03 | 58.64 | — | — | 84.11 | 58.07 | 67.62 | 73.48 | 44.79 | 54.69 | 24.63 | |
| 78B-BaseBase Model=InternVL32026.03 | 58.52 | 18.32 | 13.89 | 86.18 | 58.23 | 68.3 | 75.69 | 44.33 | 55.01 | 27.85 | |
| 7B-BaseBase Model=Qwen2-VL2026.03 | 57.96 | 29.39 | 16.59 | 83.69 | 56.79 | 66.47 | 69.96 | 43.27 | 52.65 | 17.57 | |
| BaseBase Model=Qwen2-VL-7B2026.03 | 57.96 | — | — | 83.69 | 56.79 | 66.47 | 69.96 | 43.27 | 52.65 | 17.57 | |
| Qwen2-VL-7B2026.03 | 57.96 | 29.39 | 16.59 | — | — | 66.47 | — | — | 52.65 | 17.57 | |
| ShareCaptioner2026.03 | 57.9 | 39.09 | 23.05 | — | — | 66.05 | — | — | 52.27 | 19.47 | |
| CIMBase Model=LLaVA1.5-7B2026.03 | 57.1 | — | — | 83.12 | 52.78 | 63.38 | 72.06 | 47.42 | 56.28 | 19.87 | |
| 8B-BaseBase Model=InternVL32026.03 | 56.95 | 12.57 | 11.66 | 85.98 | 55.39 | 66.08 | 74.75 | 43.19 | 53.72 | 25.11 | |
| BaseBase Model=InternVL3-8B2026.03 | 56.95 | — | — | 85.98 | 55.39 | 66.08 | 74.75 | 43.19 | 53.72 | 25.11 | |
| 7B-BaseBase Model=Qwen2.5-VL2026.03 | 55.89 | 22.68 | 14.67 | 84.64 | 54.96 | 65.06 | 72.15 | 42.13 | 52.27 | 24.35 | |
| BaseBase Model=Qwen2.5-VL-7B2026.03 | 55.89 | — | — | 84.64 | 54.96 | 65.06 | 72.15 | 42.13 | 52.27 | 24.35 | |
| BaseBase Model=LLaVA1.5-7B2026.03 | 50.72 | — | — | 81.13 | 48.37 | 59.41 | 64.12 | 39.78 | 48.01 | 9.19 |